A method for evaluating the residual life of a high-temperature structure

Abstract

The present application belongs to the technical field of condition-based maintenance support, and particularly relates to a high-temperature structure residual life evaluation method, comprising extracting a service component and sampling an in-situ test piece of a dangerous position; determining a test load according to multi-physical field simulation and equal damage theory, carrying out an in-situ life test under vacuum conditions, and obtaining surface displacement field, microstructure morphology, macroscopic strain rate and life data in the test process; taking the above data as input, training a life prediction model through machine learning; sampling batch components that need to be evaluated for residual life, and evaluating the residual life by using the trained life prediction model. The high-temperature structure residual life evaluation method directly samples the high-temperature structure after service, combines in-situ testing under vacuum environment and machine learning technology, realizes accurate prediction of the residual life of the high-temperature structure, and provides a scientific basis for the change of equipment from time-based maintenance to condition-based maintenance.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of equipment on-condition maintenance support, and particularly relates to a high-temperature structure residual life evaluation method. BACKGROUND

[0002] In recent years, high-temperature structures such as gas turbine blades, aero-engine turbine blades and nuclear reactor components play a crucial role in the fields of aerospace, energy, metallurgy and chemical industry. These key components are often in extreme service environments such as high temperature, high pressure and high stress, and their structural integrity and residual life are directly related to the safety and reliability of equipment. Therefore, how to accurately evaluate the residual life of high-temperature structures has become an important technical issue to ensure the safe operation of equipment.

[0003] On the other hand, the replacement cost of high-temperature structures is usually high. With the arrival of the overhaul period of domestic equipment in the future, accurate residual life evaluation can extend the service life of high-value structures and realize the transformation of equipment from "timed maintenance" to "on-condition maintenance", which has great economic benefits.

[0004] In the field of residual life evaluation of high-temperature structures, existing technologies mainly follow three technical paths, but each has certain limitations.

[0005] The first technical path is to record the data of the structure during service and combine it with the pre-established life model to evaluate the residual life. For example, the Chinese patent with the publication number CN118036389A discloses a method for predicting the residual life of an aero-engine blade and an aero-engine, and the Chinese patent with the publication number CN118013814A discloses a high-temperature air-cooled turbine blade life prediction method. These documents are based on existing life models and load history to determine whether the structure has reached the design life. However, this method does not refer to the current material state of the actual structure when evaluating the residual life, so it may not fully exploit the service potential of the structure, resulting in conservative evaluation results.

[0006] The second technical path is to simulate the material state after service in the laboratory and carry out corresponding tests to predict the residual service life of high-temperature structures. The Chinese patent with the publication number CN117828778A discloses a Larson-Miller curve creep life prediction method based on long-term thermal exposure test. This scheme artificially prepares deteriorated materials and conducts life tests to obtain a life model. However, a significant disadvantage of this method is that it cannot truly reflect the material and damage state of the actual structure matrix, because there may be a large difference between laboratory conditions and actual service environment.

[0007] The third technical path is to sample from the real structure, and to judge the remaining life by analyzing the microstructure morphology of the sample. For example, a nickel-based single crystal turbine blade creep remaining life prediction method is disclosed in Chinese patent publication No. CN111008495A, which obtains the microstructure morphology by sampling, and predicts the remaining life according to the established life model. However, the microstructure morphology of the material does not have a one-to-one correspondence with its historical service process, so this method may not accurately describe the actual damage state of the material, thereby affecting the accuracy of the remaining life assessment.

[0008] From the perspective of macro failure mechanism, the failure of high-temperature structure is mainly caused by surface damage of the component and failure of the material matrix inside the structure. In the on-condition maintenance system, the surface damage of the component can be monitored and excluded by regular in-situ non-destructive testing. However, the material matrix inside the structure is subjected to the combined effects of high temperature, mechanical stress, oxidation corrosion and other factors during service, and its failure mechanism is quite complex, often accompanied by microstructure evolution and material performance degradation, which cannot be predicted by the standard part life results.

[0009] Therefore, the remaining life of the actual structure cannot be accurately predicted by the life test results of the standard parts, which highlights the deficiencies of the existing technology in the remaining life assessment of high-temperature structures. In addition, the existing test methods also have some obvious limitations. When the material is processed into a small-sized test piece and subjected to high-temperature test under atmospheric conditions, the thin-wall effect caused by oxidation will significantly reduce the life of the material, thus failing to truly reflect the actual remaining service capability of the material matrix. On the other hand, the existing test methods have clear requirements for the size of the test piece, especially for high-temperature tests, which limits the possibility of sampling test on the characteristic parts of the component. Moreover, the commonly used test fixture with knurls mainly relies on friction to fix the test piece, which makes the installation process more difficult and may affect the accuracy of the test results. In addition, the microstructure and surface damage features of the material usually exist in the micron / sub-micron scale, and the conventional observation methods cannot observe the changes of these features in real time, which limits our understanding of the material failure process. Therefore, the existing high-temperature structure remaining life assessment methods have many deficiencies and cannot truly and accurately assess the remaining service capability of the structure. In order to meet the needs of the industry for on-condition maintenance, it is urgent to provide a new remaining life assessment method and complete process to overcome the limitations of the existing technology and improve the accuracy and reliability of the remaining life assessment. SUMMARY

[0010] Therefore, the present application aims to solve the problems of insufficient accuracy and technical limitations in the field of high-temperature structure residual life evaluation. The present application discloses a high-temperature structure residual life evaluation method, which obtains multiple data through in-situ test piece sampling and in-situ life test under vacuum conditions, and uses machine learning to train a life prediction model, so as to more truly and accurately evaluate the residual service capability of high-temperature structures and improve the accuracy and reliability of the evaluation.

[0011] To achieve the above-mentioned purpose, the technical scheme of the present application is as follows:

[0012] A high-temperature structure residual life evaluation method, comprising:

[0013] S1: According to the in-situ test loading scheme of the micro-sized sampling test piece, the service component is extracted and the in-situ test piece sampling of the dangerous position is carried out;

[0014] S2: According to the multi-physical field simulation and equal damage theory, the test load is determined, the in-situ life test under vacuum conditions is carried out, and the surface displacement field, microstructure morphology, macroscopic strain rate and life data in the test process are obtained by non-contact mechanical method;

[0015] S3: The multiple surface displacement fields, microstructure morphologies, macroscopic strain rates and life data obtained by in-situ test are taken as inputs, and a life prediction model is trained by machine learning;

[0016] S4: Sampling is carried out on the batch components whose residual life needs to be evaluated, and the residual life is evaluated by the life prediction model.

[0017] Further, in step S1, according to the historical failure record and non-destructive testing record, combined with full three-dimensional multi-physical field simulation, the dangerous position is judged, the sampling position and direction of the micro in-situ test piece are determined, and the sampling scheme is determined according to the local structure geometric characteristics.

[0018] Further, in step S1, the fixture is designed and checked according to the in-situ test loading scheme, the geometric structure of the micro in-situ test piece is formed, then the batch components whose residual life needs to be determined are sampled and sampled, the deformation layer of the material is removed, and the microstructure of the sampling piece is checked.

[0019] Further, the loading scheme comprises two asymmetric buckle fixtures and a heating table, the asymmetric buckle fixtures are arranged on both sides of the length direction of the test piece, the asymmetric buckle fixtures are used to provide tensile load in the axial direction of the test piece, and the heating table is used to heat the test surface and limit the displacement of the test surface.

[0020] Further, the heating table comprises:

[0021] A heating ceramic is used to generate heat and heat the test section of the in-situ test piece.

[0022] thermocouple, two of which are provided, one of which is connected to the heating ceramic for forming a negative feedback channel to ensure accurate application and maintenance of temperature, and the other of which is on the upper surface of the test piece for determining whether the test site has reached thermal equilibrium;

[0023] graphite sheet for transferring heat to the upper surface of the test piece;

[0024] elastic base provided in the clamping groove of the asymmetric clamping fixture for limiting the surface normal displacement of the test piece;

[0025] heat insulation baffle for protecting the electron gun of the scanning electron microscope in high temperature tests, which is cooled by a serpentine water cooling channel and horizontally moved by a servo motor.

[0026] Further, in step S2, the constitutive relation of the local material after service is obtained by using a non-contact mechanical testing method, and the full three-dimensional multi-physical field simulation result is dynamically corrected based on the constitutive relation of the material after service, and the in-situ test load spectrum is determined based on the equal damage principle, and the test load spectrum is any one of the creep damage design load spectrum, the creep-low cycle interactive load spectrum, and the variable load creep load spectrum.

[0027] Further, the determination of the in-situ test load spectrum based on the equal damage principle includes the following steps:

[0028] According to the actual service load spectrum, the creep duration and the low cycle load cycle are extracted;

[0029] In combination with the load of the characteristic part under each working condition determined by the multi-physical field simulation in step S1, the creep load and the low cycle load cycle of the characteristic part are determined;

[0030] When it is detected that the damage of the test piece under the creep load is greater than the low cycle fatigue damage by an order of magnitude, only the creep damage design load spectrum is considered, or when it is detected that the low cycle fatigue damage of the test piece is greater than the damage under the creep load by an order of magnitude, only the low cycle fatigue damage design load spectrum is considered, otherwise, according to the damage proportion of the two, the creep-low cycle interactive load spectrum is designed.

[0031] Further, in step S2, the non-contact mechanical method includes digital image correlation method or moire method, and in the process of in-situ life test on the sampling piece, the strain-time curve, the surface displacement distribution of the test piece are obtained by the non-contact mechanical testing method, and the microstructure information is extracted by using the digital image processing scheme.

[0032] Further, in step S2, a scanning electron microscope with high magnification and large depth of field is combined with the test to observe the in-situ microstructure and surface damage evolution of the test piece, and the scanning electron microscope adopts a control electron beam point-by-point scanning imaging.

[0033] Further, in step S3, a hybrid model using a combination of long short-term memory network and convolutional neural network is used for data processing and feature extraction to identify displacement field features, microstructure features, creep rate and the implicit relationship between the residual life, wherein the long short-term memory network can capture the dependency relationship in the time series, and the convolutional neural network can automatically extract local spatial features.

[0034] Compared with the prior art, the high-temperature structure residual life evaluation method has the following advantages:

[0035] 1. The high-temperature structure residual life evaluation method can accurately locate the dangerous parts of the high-temperature structure by combining multi-physical field simulation analysis, directly sample the dangerous parts of the service parts, and realize real-time collection of key data such as sample surface displacement field, microstructure morphology and macroscopic strain rate under non-contact mechanical detection in a vacuum environment, and accurately determine the test load by combining multi-physical field simulation and equal damage theory, and on this basis, a large number of in-situ test data are analyzed and mined by a machine learning model to realize accurate prediction of the residual life of the high-temperature structure, and provide a scientific basis for the transformation from time-based maintenance to condition-based maintenance of high-temperature equipment. 2. The high-temperature structure residual life evaluation method is simple to operate, low in cost, high in data acquisition and processing efficiency, and can be widely used for residual life prediction of high-temperature service structures such as gas turbine blades, aircraft engine turbine blades and nuclear reactor components. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a logic block diagram of the high-temperature structure residual life evaluation method according to the embodiment of the application;

[0037] Figure 2 is a schematic diagram of the asymmetric bayonet structure according to the embodiment of the application;

[0038] Figure 3 is a general schematic diagram of the test piece loading scheme according to the embodiment of the application;

[0039] Figure 4 is an example of continuously obtaining the surface damage and speckle images of the test piece under a scanning electron microscope;

[0040] The marks in the figure are: 100-asymmetric bayonet clamp; 200-test piece; 300-heating ceramic; 400-graphite sheet; 500-thermocouple; 600-heat insulation baffle. DETAILED DESCRIPTION

[0041] With reference to the drawings and the embodiments described herein, it will be understood that the drawings and embodiments are illustrative of only a few of the embodiments of the present application and are not therefore to be considered limiting of its scope, for the application is not limited to the embodiments illustrated in the description below.

[0042] In the description of the present application, it should be understood that the terms used herein are for the purpose of describing specific embodiments and are not intended to limit the example embodiments of the present application. For the purpose of clarity, the dimensions of the various parts shown in the drawings are not drawn to scale. Techniques, methods, and apparatus known to those of ordinary skill are not discussed in detail because they would be understood that the techniques, methods, and apparatus are part of the prior art and should be understood to be incorporated into the disclosure in order to fully enable the present application. In all examples shown and discussed herein, any specific values are to be interpreted as illustrative only and not limiting. Thus, other examples of the example embodiments can have different values. It is noted that like numbers and letters on the figures identify like parts throughout the disclosure, and thus, once defined, do not need to be further discussed.

[0043] It should be noted that the terms "first", "second", and the like, used in the description and in the claims of the present application are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the use of these terms herein is merely for distinguishing between similar elements and is not meant to limit the scope of embodiments of the present application to a specific order or sequence. Also, the terms "first", "second", and the like, used in the description and in the claims of the present application are used for distinguishing between similar elements and do not imply a particular order or sequence, unless otherwise specified herein. It is to be understood that the terms "comprising", "including", "containing", or "having" and variations thereof, when used herein, are meant to encompass the items listed thereafter, and equivalents thereof as well as additional items, unless otherwise specified herein.

[0044] It should be noted that the terms "front", "rear", "upper", "lower", "left", "right", "horizontal", "vertical", and "top", "bottom", and the like, used in the description and in the claims of the present application are used for describing relative positions on the drawings and are not meant to be limiting as to the application of the present application. These terms are used for convenience only and are not to be construed as limiting the scope of the present application to only those embodiments where one part is positioned on top of another part, for example, as shown in the drawings. The terms "front", "rear", "upper", "lower", "left", "right", "horizontal", "vertical", and "top", "bottom", and the like, used in the description and in the claims of the present application are used for describing relative positions on the drawings and are not meant to be limiting as to the application of the present application. These terms are used for convenience only and are not to be construed as limiting the scope of the present application to only those embodiments where one part is positioned on top of another part, for example, as shown in the drawings. The terms "front", "rear", "upper", "lower", "left", "right", "horizontal", "vertical", and "top", "bottom", and the like, used in the description and in the claims of the present application are used for describing relative positions on the drawings and are not meant to be limiting as to the application of the present application. These terms are used for convenience only and are not to be construed as limiting the scope of the present application to only those embodiments where one part is positioned on top of another part, for example, as shown in the drawings.

[0045] It should be noted that in this application, the terms "comprising", "including", or any other variant thereof are intended to cover non-exclusive inclusions, so that processes, methods, articles, or devices that include a series of elements not only include those elements, but also include other elements not explicitly listed, or inherent to such processes, methods, articles, or devices. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article, or device including the element. In addition, it should be pointed out that the scope of the methods and devices in the embodiments of the present application is not limited to the order of performing the functions shown or discussed, but can also include performing the functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method can be performed in an order different from that described, and various steps can also be added, omitted, or combined. In addition, the features described with reference to certain examples can be combined in other examples.

[0046] Applicants have found in the current technical practice that taking small-size test pieces from service structures and conducting life tests is considered as a method that can effectively reflect the real residual service capacity and residual life of the material matrix at a specific position. The advantage of this method is that it directly evaluates the actual material state during service, which has high accuracy. On the other hand, with the continuous progress of observation means, it is now possible to record the damage and microstructure state of the material in real time during the test, which provides strong support for in-depth analysis of the failure mechanism of the material. In addition, the rapid development of machine learning makes it possible to mine implicit life laws from complex test phenomena, which further improves the accuracy and reliability of residual life assessment.

[0047] Based on this, the present application discloses a high-temperature structure residual life assessment method, comprising:

[0048] S1: According to the in-situ test loading scheme of the micro-size sampling test piece, the service component is extracted and the in-situ test piece sampling of the dangerous position is conducted;

[0049] S2: According to the multi-physical field simulation and equal damage theory, the test load is determined, the in-situ life test under vacuum condition is carried out, and the surface displacement field, microstructure morphology, macroscopic strain rate and life data in the test process are obtained by non-contact mechanical method;

[0050] S3: The multiple surface displacement fields, microstructure morphologies, macroscopic strain rates and life data obtained by the in-situ test are taken as inputs, and a life prediction model is trained by machine learning;

[0051] S4: Sampling is conducted on the batch components whose residual life needs to be assessed, and the residual life is assessed by the life prediction model.

[0052] The method for assessing the remaining life of high-temperature structures disclosed in this application, such as Figure 1 As shown, by selecting representative hazardous areas in service components for sampling based on the in-situ testing loading scheme of small-sized sampled specimens, key areas are precisely located according to the local geometric features and actual service conditions of the components during the sampling process. By directly sampling from the high-temperature structure after service, the deterioration effect of the service environment on the material matrix can be accurately reflected, avoiding the problem of mismatch between laboratory simulation methods and actual service conditions. Then, the sampled specimens are subjected to in-situ testing in a vacuum environment, effectively avoiding the oxidation effect of small-sized samples at high temperatures, making the test results closer to the actual performance of the material. During the test, non-contact mechanical testing methods are used to obtain dynamic data such as surface displacement distribution, microstructure morphology, and macroscopic strain rate. The large amount of data collected is preprocessed and used as input. This method utilizes machine learning to train a life prediction model, enabling data-driven intelligent analysis and prediction. Based on experimental data reflecting the actual service condition, it effectively overcomes the localized biases caused by sample size limitations, experimental environment differences, and processing damage inherent in traditional assessment methods. It achieves seamless integration from precise collection of service-ready samples and efficient acquisition of dynamic data to the training and prediction of the intelligent model, providing an accurate and reliable basis for assessing the remaining life of subsequent batches of components. The overall process fully leverages advanced simulation technology, in-situ testing methods, and intelligent data analysis to comprehensively quantify the performance degradation and damage evolution patterns of high-temperature structural materials, providing real-time and effective data support and scientific guidance for equipment maintenance decisions. It is particularly suitable for the rapid assessment of large-scale service equipment.

[0053] The high-temperature structure remaining life assessment method described in this invention utilizes in-situ testing of small-sized specimens at critical locations of service components. Non-contact mechanical testing under vacuum conditions enables real-time acquisition of key data such as the surface displacement field, microstructure morphology, and macroscopic strain rate of the specimens. Combined with multiphysics simulation and equal damage theory, the test load is accurately determined. Based on this, a machine learning model is used to deeply analyze a large amount of in-situ test data to uncover the inherent laws of material degradation and life evolution, thereby achieving accurate prediction of the remaining life of high-temperature structures. This method effectively solves the problems of sample processing-introduced damage, discrepancies between laboratory simulation and actual service environment, and limitations in data acquisition inherent in traditional life assessment techniques. It not only accurately reflects the actual degradation state of the material matrix under high temperature, high pressure, and complex load conditions, but also captures the material damage evolution process in a timely manner during dynamic monitoring, providing a scientific basis for the transformation of equipment maintenance from scheduled maintenance to condition-based maintenance. The process is simple, data processing is efficient, model prediction is accurate, and the applicability is wide, enabling large-scale assessment of the remaining life of high-temperature components.

[0054] As a preferred example of the present application, in step S1, according to the historical failure records and non-destructive testing records, combined with full three-dimensional multi-physical field simulation, the dangerous position is judged, the sampling position and direction of the micro in-situ test piece are determined, and the sampling scheme is determined according to the local structure geometric characteristics. In the specific example of the present application, according to the historical failure records and non-destructive testing records, combined with full three-dimensional multi-physical field simulation, the service structure is comprehensively analyzed, the simulation results are used to determine the areas where stress concentration, temperature anomaly and load superposition exist in the structure, and then the representative dangerous positions are judged, and the optimal sampling position and sampling direction of the micro in-situ test piece are determined according to the geometric characteristics of the local structure such as thickness, curvature, boundary and connection mode, etc. The sampling scheme is further optimized in angle and direction in combination with the first principal stress direction and the actual load action direction, that is, in the example of the present application, the sampling direction and the test piece plane normal are determined according to the local first principal stress direction of the structure dangerous position, the local geometric characteristics are measured and compared through finite element simulation, so as to determine the size, shape and sampling angle of the sample, so as to ensure that the collected sample can truly reflect the distribution of internal damage and degradation of the material. In the sampling process, the optimized processing technology is used to reduce the additional damage caused by mechanical cutting or polishing, and the multi-physical field simulation results and the field detection data are verified with each other, so as to realize the accurate formulation of the sampling scheme of the dangerous position.

[0055] The high-temperature structure residual life evaluation method disclosed in the present application realizes the comprehensive capture of the internal damage and degradation process of the material through the accurate identification of the dangerous position of the service component and the optimized sampling of the micro sample, provides reliable data support for the equipment condition-based maintenance, not only provides a solid data foundation for obtaining real surface displacement field, microstructure morphology and macroscopic strain rate data through non-contact mechanical method in subsequent in-situ test, but also lays a foundation for high-quality input data collection of machine learning model, and significantly improves the accuracy and real-time performance of residual life prediction.

[0056] As a preferred example of the present application, in step S1, the jig is designed and checked according to the in-situ test loading scheme, the geometric structure of the micro in-situ test piece is formed, then a sample is sampled and sampled from the batch components to be determined residual life, the deformation layer of the material is removed, and the microstructure of the sampling piece is checked. In the example of the present application, in step S1, the special jig is designed and checked according to the in-situ test loading scheme to form the accurate geometric structure of the micro in-situ test piece, and a representative sample is randomly taken from the batch components to be determined residual life for processing to remove the deformation layer caused by mechanical cutting and polishing, and the microstructure of the sampling piece is checked in detail to ensure that the sample truly reflects the service state of the material.

[0057] The setting significantly improves the accuracy and real-time performance of the high-temperature structure residual life assessment by optimizing the clamp loading scheme and fine sampling processing technology, accurately determines the dangerous positions existing in the service components by using full three-dimensional multi-physical field simulation combined with historical failure records and nondestructive testing data, thereby guiding the accurate sampling of the micro in-situ test piece, and ensures the authenticity of the microstructure of the test sample by removing the deformation layer introduced in the processing process. Not only solves the data distortion problem caused by small sample size, processing damage and atmospheric oxidation in traditional sampling test, but also effectively reduces the dispersion of the evaluation results, and provides scientific and reliable data support for the transformation of equipment from time-based maintenance to condition-based maintenance.

[0058] As a preferred example of the present application, when the microstructure of the sampling piece is inspected in step S1, for the advanced gas turbine blade using nickel-based single crystal high-temperature alloy, the inspection includes quantitative and qualitative analysis of the volume fraction, size and morphology of the strengthening phase (γ'), and the presence of microvoids and TCP phases. As a specific example of the present application, the inspection of the microstructure of the sampling piece is determined according to the material characteristics, for the mainstream nickel-based single crystal high-temperature alloy of the advanced gas turbine blade, the microstructure mainly includes the volume fraction, size and morphology of the strengthening phase (γ'), microvoids, TCP phases, etc. High-resolution scanning electron microscope and digital image processing are used to obtain the microstructure data of the sampling piece, to ensure that the collected microstructure information can truly reflect the structural degradation and performance degradation of the material during service, significantly improving the accurate reflection ability of the actual service state of the material, and dynamically correcting the material constitutive relationship combined with the full three-dimensional multi-physical field simulation results, effectively solving the data distortion problem caused by processing damage and environmental interference in the traditional sampling and detection process.

[0059] As a specific example of the present application, the loading scheme includes two asymmetric buckle clamps 100 and a heating table, the asymmetric buckle clamps 100 are arranged on both sides of the length direction of the test piece 200, the asymmetric buckle clamps 100 provide the tensile load in the axial direction of the test piece 200, and the heating table is used to heat the test site and limit the displacement of the test surface normal. In the example of the present application, the asymmetric buckle clamps 100 are used to facilitate the clamping of the test piece 200, wherein the design scheme of the asymmetric buckle clamps 100 is as follows:

[0060] The clamping groove in one clamp can closely fit the entire clamping section of one end of the in-situ test piece 200, and the clamping groove in the other clamp is only used to prevent displacement in the tensile direction of the other end of the in-situ test piece 200.

[0061] As Figure 2As shown, the asymmetric clamping fixture 100 described in the present application, the displacement perpendicular to the surface of the test piece 200 is limited by the clamping groove bottom and the heating platform, if necessary, the flat groove of the clamping groove can be changed to dovetail groove. This design reduces the number of clamps and installation steps, and since the test piece 200 is controlled by the groove surface during installation, the process of manual centering is avoided, the installation precision of the test piece can be better controlled, and the dispersion caused by the test process is reduced.

[0062] In the example of the present application, the heating platform comprises:

[0063] The heating ceramic 300 is used to generate heat and heat the test piece 200 in situ;

[0064] The thermocouple 500 is provided with two, one of which is connected to the heating ceramic 300, which is used to form a negative feedback channel to ensure accurate application and maintenance of temperature, and the other is on the upper surface of the test piece 200, which is used to judge whether the test piece 200 reaches thermal equilibrium;

[0065] The graphite sheet 400 is used to transfer heat to the upper surface of the test piece 200;

[0066] The elastic base is arranged in the clamping groove of the asymmetric clamping fixture 100, which is used to limit the surface normal displacement of the test piece 200;

[0067] The heat insulation baffle 600 is used to protect the scanning electron microscope electron gun in high temperature test, which is cooled by a serpentine water cooling channel and moves horizontally by a servo motor.

[0068] The loading scheme in step S1 of the present application is composed of two asymmetric clamping fixtures 100 and a heating table, wherein the asymmetric clamping fixtures 100 are arranged on both sides of the test piece 200 respectively, the clamping groove of one side fixture can closely fit the whole clamping section of one end of the test piece 200 to ensure the stability of the test sample when the axial tensile load is applied, and the clamping groove of the other side fixture is only used to prevent the displacement of the test sample in the tensile direction, and the displacement of the test sample perpendicular to the surface thereof is limited through the cooperation of the clamping groove bottom and the heating table, and the clamping centering accuracy is further improved by changing the flat groove into a dovetail groove if necessary, and the heating table is composed of a heating ceramic 300, two thermocouples 500, a graphite sheet 400, an elastic base and a heat insulation baffle 600, wherein the heating ceramic 300 generates uniform heat through voltage control to heat the test part of the test piece 200, one thermocouple 500 is connected with the heating ceramic 300 to form a negative feedback channel to ensure accurate temperature application, and the other thermocouple 500 is arranged on the surface of the test piece 200 for real-time monitoring of the heat balance state, and the temperature is closed-loop controlled through the two thermocouples; the graphite sheet 400 is used for heat transfer, and can be appropriately trimmed according to the part that needs to be heated, such as when the total length of the test piece is less than the size of the heating table, the size of the graphite sheet can be reduced to reduce the heating power of the clamping end and improve the design margin of the clamping end; the elastic base is used to limit the normal displacement of the surface of the test piece 200, and the balance position is located on the symmetry plane of the loading equipment to ensure the overall balance of the loading equipment, and the heat insulation baffle 600 is horizontally moved by a servo motor through a serpentine water cooling channel to protect the electron gun of the scanning electron microscope, so that the surface displacement field, microstructure morphology and macroscopic strain rate and other key data of the test piece 200 in the test process are collected in real time under vacuum conditions by the non-contact mechanical test method, and a large amount of collected data is preprocessed and input into the machine learning model for life prediction. In the specific example of the present application, the width-thickness ratio of the test part of the micro-size sampling test piece 200 in step S1 is greater than 2, and in order to avoid thermal deformation caused by mechanical processing during sampling, the thickness of the micro-size sampling test piece 200 should not be less than 0.6mm; the strength of the test piece 200 and the fixture as a whole should be checked by finite element simulation. For high temperature test, the strength requirement of the contact surface near the test piece and the fixture can be appropriately relaxed due to the rapid temperature drop outside the test part.

[0069] By adopting the optimized loading scheme composed of two asymmetric clamping fixtures 100 and a heating table, the present application effectively improves the test sample clamping accuracy and loading stability, and the whole process realizes seamless connection from test sample preparation, accurate loading, dynamic data acquisition to data-driven intelligent life prediction, thereby providing scientific and reliable data support for equipment condition-based maintenance.

[0070] As a preferred example of the present application, the constitutive relation of the local material after service is obtained by using a non-contact mechanical testing method in step S2, and the full three-dimensional multi-physical field simulation result is dynamically corrected based on the constitutive relation of the material after service, and the in-situ test load spectrum is determined based on the equal damage principle, and the test load spectrum is any one of the creep damage design load spectrum, the creep-low cycle interactive load spectrum, and the variable load creep load spectrum. As a specific example of the present application, the in-situ test load spectrum is determined based on the equal damage principle, which includes the following steps:

[0071] According to the actual service load spectrum, the creep duration and the low cycle load cycle are extracted;

[0072] Combined with the load of the characteristic position under each working condition determined by the multi-physical field simulation in step S1, the creep load and the low cycle load cycle of the characteristic position are determined;

[0073] When it is detected that the damage of the test piece under the creep load is greater than the low cycle fatigue damage by an order of magnitude, only the creep damage design load spectrum is considered, or when it is detected that the low cycle fatigue damage of the test piece is greater than the damage under the creep load by an order of magnitude, only the low cycle fatigue damage design load spectrum is considered, otherwise, according to the damage proportion of the two, the creep-low cycle interactive load spectrum is designed.

[0074] In step S2 of the present application, the load spectrum is determined by multi-physical field simulation and equal damage theory, wherein the multi-physical field simulation determines the size of the load, and the equal damage theory determines the form of the load. According to the load characteristics of the service component, it can be designed as creep, creep-low cycle interaction, variable load creep, etc., and if necessary, an accelerated test spectrum can be designed.

[0075] By adopting the non-contact mechanical testing method in step S2 to obtain the constitutive relationship of the local material after service in real time, and using the obtained constitutive data to correct the full three-dimensional multi-physical field simulation results to accurately reflect the real performance change of the material under the action of high temperature, high pressure and complex load, and according to the creep time length and low cycle load cycle data extracted from the actual service load spectrum, combined with the load information of each feature position determined by the multi-physical field simulation in step S1, the damage ratio of the material under the conditions of creep and low cycle fatigue is calculated by using the equal damage principle, and the in-situ test load spectrum which can reflect the creep damage and consider the low cycle fatigue interaction is designed. When the creep damage is significantly greater than the low cycle fatigue damage, only the creep damage is considered, and vice versa, otherwise the creep and low cycle fatigue interaction load spectrum is designed according to the damage ratio. Through the organic combination of non-contact mechanical testing method, full three-dimensional multi-physical field simulation, equal damage principle and deep learning technology, the real-time acquisition and accurate correction of the constitutive relationship of the local material after service are realized, which ensures that the design of the in-situ test load spectrum can fully reflect the actual damage accumulation of the material under the action of high temperature, high pressure and complex load, thereby significantly improving the accuracy and real-time performance of the remaining life prediction, and providing reliable technical support for the transformation of equipment from time-based maintenance to condition-based maintenance.

[0076] As a preferred example of the present application, in step S2, if the actual load spectrum of the service component structure is complex, multiple groups of in-situ test loads are set. In the example of the present application, if the load data of the service structure under different operating conditions is comprehensively collected and analyzed, multiple load cycles, temperature fluctuations and load interaction effects experienced by the structure in the actual working condition are identified, and the duration, amplitude and frequency of each type of load are finely divided by combining full three-dimensional multi-physical field simulation, historical failure records and on-site non-destructive testing data, thereby determining the complex load spectrum of the structure during service. The complex load spectrum not only includes long-term sustained creep load, but also covers short-term low cycle fatigue and mixed load effects formed by the interaction of the two, for example, the complex load spectrum includes a combination state composed of load cycles, temperature fluctuations and load interaction effects under multiple operating conditions. In step S2, multiple groups of in-situ test loads are set, each load spectrum is independently designed for pure creep load, low cycle fatigue and creep and low cycle fatigue interaction, and for structures with intermittent overload, severe load fluctuations and multiple working conditions, the load duration, amplitude and frequency of each working condition are finely divided to achieve comprehensive coverage of each type of load state under complex service environment, thereby avoiding the deviation of the remaining life prediction caused by using a single load spectrum for extrapolation prediction.

[0077] As a preferred example of the present application, in step S2, the non-contact mechanics method includes digital image correlation method or moire method, and during the in-situ life test of the sample, the strain-time curve, the surface displacement distribution of the sample and the microstructure information are obtained by the non-contact mechanics test method respectively. In the specific example of the present application, the non-contact mechanics test method used in step S2 is limited to digital image correlation method or moire method, which uses a high-resolution digital camera to continuously collect image sequences generated by the sample during the in-situ life test, and through advanced image processing algorithms, the sample surface is tracked and quantitatively analyzed to obtain the strain-time curve and the surface displacement distribution of the sample in real time, and the microstructure information is extracted through the digital image processing scheme to realize the dynamic monitoring of the surface damage and microstructure evolution of the sample. The test method avoids the interference and installation error introduced by the traditional contact sensor, and can accurately reflect the real degradation state of the material under high temperature, high pressure and complex load, thereby significantly improving the accuracy and repeatability of data acquisition.

[0078] As a preferred example of the present application, in step S2, the scanning electron microscope with high magnification and large depth of field is combined with the test to observe the in-situ microstructure and surface damage evolution of the sample, and the scanning electron microscope adopts control electron beam point-by-point scanning imaging. In the example of the present application, in order to realize the accurate in-situ observation of the microstructure and surface damage evolution of the high-temperature structural material during service, the scanning electron microscope with high magnification and large depth of field is combined with the in-situ test system in step S2, which can realize the in-situ observation of the microstructure and surface damage evolution during the experiment, and the cost is relatively low. The scanning electron microscope uses its excellent imaging resolution and wide depth of field to continuously collect sample surface images by controlling electron beam point-by-point scanning during the test. Although the shooting speed is slow, under the service load conditions of high-temperature components, the stress level of the sample is low and the plastic deformation of the material during imaging can be ignored, thereby avoiding the adverse effects of slow imaging speed.

[0079] By combining the scanning electron microscope with high magnification and large depth of field with the in-situ test system in step S2, the present application not only fully utilizes the advantages of the scanning electron microscope in capturing microstructure details, but also effectively avoids the problem of slow imaging speed of electron beam point-by-point scanning by taking advantage of the small deformation of the material under low stress service load, thereby realizing real-time, continuous and non-contact collection of the sample surface displacement field, microstructure evolution and surface damage. The high-quality image data collected after standardization processing and automatic digital image analysis can truly reflect the degradation state of the material under high temperature, high pressure and complex load conditions, and provide high-precision, stable and reliable input data for accurate prediction of remaining life based on deep learning model.

[0080] As a preferred example of the present application, in step S3, a hybrid model combining the use of long short-term memory network (LSTM) and convolutional neural network (CNN) is used for data processing and feature extraction to identify displacement field features, microstructure features, creep rate and their implicit relationship with residual life, wherein LSTM can capture the dependency relationship in time series, and CNN can automatically extract local spatial features. In the example of the present application, in step S3, a hybrid model combining the use of long short-term memory network (LSTM) and convolutional neural network (CNN) is used to realize real-time, non-contact continuous collection of specimen surface displacement field, macroscopic strain rate and microstructure morphology data collected by digital image correlation method or cloud pattern method during in-situ test, and the displacement data is obtained by comparing the process images with the initial no-load state image as a reference and then standardized, while the volume fraction, size and morphology of the γ / γ' two-phase in the nickel-based single crystal high-temperature alloy are automatically extracted by using the open source digital image processing program. In the hybrid model constructed by tensorflow.keras, the CNN automatically extracts local spatial features through multiple convolutional layers and pooling layers, and the LSTM captures the dependency relationship in time series data through multiple LSTM layers. Subsequently, the high-dimensional features output by each layer are fused through a fully connected layer, and finally the model is trained and cross-validated using the Adam optimizer and mean square error loss function, forming a comprehensive description of the degradation behavior of the material under the action of high temperature, high pressure and complex load. Then, the trained deep learning model is used to accurately predict the residual life of the material, so as to accurately predict the residual life of the subsequent batch structure.

[0081] The method of training the life prediction model by machine learning in the present application not only overcomes the prediction deviation caused by the discontinuous sample data collection, insufficient feature extraction and single model limitations in traditional methods, but also accurately reflects the material degradation law under complex service conditions through the trained model, providing a scientific, reliable and real-time decision basis for equipment maintenance and on-demand maintenance.

[0082] In the example of the present application, the input of the life model in step S3 includes surface displacement distribution, microstructure morphology, macroscopic strain rate and residual life at a certain state during the test, wherein the residual life = total test life - current test acquisition time.

[0083] As a preferred example of the present application, in step S4, sampling is performed based on statistical principles to ensure that the selected samples can represent the service state of the overall batch components, and the residual life evaluation results in step S4 can be used to adjust the maintenance plan of subsequent service components to realize on-demand maintenance.

[0084] The high-temperature structure residual life evaluation method disclosed in the application can accurately locate and optimize sampling of dangerous positions such as stress concentration, temperature anomaly and load superposition in the service component by taking small-size samples from the service structure and performing in-situ life test, and can remove the deformation layer introduced in the mechanical cutting, grinding and electrolytic polishing process by using advanced processing technology to ensure that the sampling piece truly reflects the service state of the material, and can collect key dynamic data such as sample surface displacement field, microstructure morphology and macroscopic strain rate in real time in a vacuum environment by using a non-contact mechanical testing method, and can realize in-situ observation of microstructure and surface damage evolution of the sample by using digital image correlation method, moire method and scanning electron microscope with high magnification and large depth of field, and can accurately extract micro features such as volume fraction, size and morphology of γ / γ' two phases in the nickel-based single crystal high-temperature alloy by using an automatic digital image processing program, and can perform standardization preprocessing, local spatial feature extraction and time series dependency analysis on the collected continuous data by using a hybrid deep learning model combining long short-term memory network and convolutional neural network, and can perform model training and cross-validation by using Adam optimizer and mean square error loss function, so as to realize dynamic correction of material degradation behavior under the action of high temperature, high pressure and complex load and accurate prediction of residual life. When facing complex service load spectrum, multiple load cycles, temperature fluctuations and load interaction effects, multiple in-situ test loads are set to simulate pure creep, low-cycle fatigue and their interaction, various service conditions are fully covered, and the prediction deviation caused by single load spectrum extrapolation is avoided, so that scientific, real-time and reliable data support is provided for the transformation of equipment from time-based maintenance to condition-based maintenance, the service life of the key components of the high-temperature structure is significantly prolonged, the maintenance cost is reduced, and the operation safety and overall efficiency of the equipment are improved. The high-temperature structure residual life evaluation method disclosed in the application can fully consider the performance degradation of the material matrix in the service environment by taking small-size sampling test pieces from the structure after service and performing life test; the load spectrum is designed according to the equal damage principle, which can reflect the residual life law of the structure under complex load conditions; the life model can be obtained by processing a large amount of data obtained by in-situ test through machine learning, which can support residual life evaluation of a large number of high-temperature structures, and is simple and convenient to operate, low in cost, high in data acquisition and processing efficiency, and can be widely used in residual life prediction of high-temperature service structures such as gas turbine blades, aero-engine turbine blades and nuclear reactor components.

[0085] The application discloses a high-temperature structure residual life evaluation method, which is specifically used for evaluating the residual life of a certain type of aero gas turbine turbine blade and comprises the following steps.

[0086] S1-1: For the main flight conditions, full three-dimensional multi-physical field simulation results of gas turbine blades are carried out to obtain the temperature, stress / strain field of the structure in the service state, and the positions with high temperature, stress / strain are preliminarily selected as dangerous positions;

[0087] S1-2: According to the detection results of the retired blade in the external field, the dangerous positions are appropriately increased, decreased or adjusted. Taking the 1 / 3 position of the blade height as an example, since the first principal stress direction is approximately parallel to the blade height direction, the sampling direction is determined to be parallel to the blade height direction;

[0088] S1-3: Combined with the local geometric model, the size of the examination section of the micro-size sample is determined to be 0.7mmx1.5mm;

[0089] S1-4: According to the size of the examination section, the clamping end of the test piece and the corresponding loading system are designed as shown in Figure 3

[0090] S1-5: A 0.8mm thick test piece is taken out from the specified position of the structure after service by slow wire electro-discharge wire cutting, and the side geometry of the test piece is machined by slow wire electro-discharge wire cutting. The surface roughness obtained by slow wire electro-discharge wire cutting is <1μm, and after manual polishing by 3000#SiC water abrasive paper, mechanical polishing by 3μm, 1μm and 0.3μm aluminum oxide polishing liquid, electrolytic polishing of the test piece by 7% perchloric acid and 93% ethanol electrolyte to remove the surface deformation layer caused by wire cutting, the size of the in-situ test piece examination section is measured, among which the thickness is measured by a precision thickness gauge, and the width is measured by a ruler in a scanning electron microscope.

[0091] S2-1: The digital image correlation method with relatively simple sample pretreatment is used to measure the surface deformation of the test piece. First, the size of the speckle particle needs to be determined according to the required field size, and the field size should be comprehensively evaluated considering the typical damage size, surface organization evolution characteristic size, etc. For the nickel-based single crystal superalloy in this example, zirconium oxide particles with an average particle size of 2μm are selected to prepare the speckle. First, the particles are uniformly distributed in the alcohol solution by ultrasonic cleaning agent, and the suspension is dropped onto the surface. The particles are randomly and uniformly distributed on the surface by rapid and uniform shaking with a glue spreader, forming a speckle pattern.

[0092] ​S2-2: Correct the image. There may be distortions in the scanning electron microscope during imaging that are not conducive to subsequent analysis, and the image distortion needs to be corrected before the formal start of the test. First, select appropriate shooting time, current size, working distance, current size, etc. Parameters can be judged by shooting multiple images, comparing noise level and data repeatability, and keeping consistent in subsequent shooting; use the method of translation test to check the measurement accuracy, and get the image by moving the test piece up and down and left and right multiple times, and the length of each translation is about 1 / 4 field of view; Use digital image correlation program to evaluate distortion level and strain measurement accuracy, and determine whether the results need to be corrected. Preferably, the strain measurement accuracy should be better than 1e-3 strain.

[0093] S2-3: Conduct in-situ tensile test, and obtain the Young's modulus and Poisson's ratio of the material after service in the sampling direction in the online elastic segment. Install the micro-sized sampling test piece into the experimental equipment, continuously control the tension and warm up to the specified temperature, and stretch to the local maximum load determined in S1-1; Get the surface picture before and after stretching, and analyze the surface displacement field through the open source program; By selecting the marker points, the average elongation in the stretching direction and the transverse direction is determined by the displacement change before and after stretching, and then the average strain is calculated, and the Young's modulus and Poisson's ratio can be calculated according to the corresponding definition formula.

[0094] S2-4: Compare the tensile properties of the material manual and the material after service. In this example, the nickel-based single crystal superalloy is targeted, and the performance degradation caused by the rafting of the strengthening phase usually has a negligible effect on the material's constitutive. If the test results differ greatly from the material manual results, the simulation results need to be recalibrated, and the reasons for the change in material constitutive parameters need to be analyzed.

[0095] S2-5: According to the load spectrum determined by the full three-dimensional multi-physical field simulation results and the equal damage principle, carry out in-situ life test. During the experiment, the surface image of the test piece is obtained regularly, as shown in Figure 4 When the scanning electron microscope obtains the image, stop the test machine beam to avoid image distortion caused by vibration.

[0096] S3-1: Process the acquired in-situ data to obtain machine learning input. In this example, open-source code Ncorr is used to analyze the surface deformation displacement field. To avoid cumulative errors, the image obtained in the initial unloaded state is used as the reference image, and the process image is used as the target image. After obtaining the displacement data, standardization processing is performed, and the image center displacement is 0. Based on the obtained surface displacement distribution-time history, the strain-time history is obtained by the method of S4 to obtain the macroscopic strain rate. In this example, since the evolution of microstructure evolution under high temperature service conditions is relatively sufficient, the evolution process and key parameters are also relatively clear. To reduce the training cost, the open-source digital image processing program ImageJ is used to process the γ / γ' two-phase morphology image to extract the γ' volume fraction, length and width size.

[0097] S3-2: A hybrid model is constructed using tensorflow.keras; the surface displacement distribution, macroscopic strain rate, microstructure morphology, test load conditions and real-time residual life obtained after the test in S3-1 are used as input; CNN is used to extract image features. Two convolutional layers and pooling layers are included, and high-dimensional features are extracted through a fully connected layer at the end; LSTM is used to process time series data. Two LSTM layers are included, and high-dimensional features are extracted through a fully connected layer at the end; image features and macroscopic strain rate, microstructure morphology and load condition features are fused, further processed through a fully connected layer, and finally the predicted residual life is output.

[0098] S3-3: The model is compiled using the Adam optimizer and the mean square error (MSE) loss function; in this example, 4 sets of test results are used as the training set, 1 set is used as the validation set, and cross-validation is used to enhance the training effect.

[0099] S4-1: The trained model is used for residual life prediction. For other batch structures, microstructure information is obtained by in-situ sampling, and short-term tests are carried out on individual pieces. After obtaining the required prediction input, the residual life of the batch structure can be quickly evaluated; new test results can also be used to improve the model.

[0100] The present application fully considers the performance degradation of the material matrix caused by the service environment by taking out small-sized sampling test pieces from the structure after service and carrying out life tests; the load spectrum is designed based on the equal damage principle, which can reflect the residual life law of the structure under complex load conditions, and the scheme has high feasibility.

[0101] The application obtains a large amount of surface damage information and microstructure evolution information based on in-situ test under a scanning electron microscope through machine learning processing, and obtains a life evaluation model, which can be used for quickly and accurately evaluating the remaining life of the remaining batches of structures. The embodiments of the application are described above in combination with the drawings, the embodiments in the application and the features in the embodiments can be combined with each other in the case of no conflict, the application is not limited to the above-mentioned specific embodiments, the above-mentioned specific embodiments are only illustrative but not limited, and a person skilled in the art can make many forms under the inspiration of the application without departing from the purpose of the application and the scope protected by the claims, which all belong to the protection of the application.

Claims

1. A method for assessing the remaining life of a high-temperature structure, characterized in that... ,include: S1: Based on the in-situ test loading scheme of the micro-sized sampling test specimen, extract the service components and perform in-situ test specimen sampling of the dangerous parts; S2: Determine the test load based on multiphysics simulation and equal damage theory, conduct in-situ life test under vacuum conditions, and obtain the surface displacement field, microstructure morphology, macroscopic strain rate and life data of the test piece during the test through non-contact mechanical methods; S3: Using multiple surface displacement fields, microstructure morphology, macroscopic strain rate and lifetime data obtained from in-situ experiments as input, a lifetime prediction model is trained through machine learning. S4: Sampling is performed on the batch of components whose remaining life needs to be assessed, and the remaining life is assessed using the life prediction model. In step S1, based on historical failure records and non-destructive testing records, combined with full three-dimensional multiphysics simulation, the dangerous parts are identified, the sampling location and direction of the micro in-situ test specimen are determined, and the sampling scheme is determined based on the local structural geometric characteristics. Among them, the sampling direction and the normal direction of the test specimen plane are determined based on the direction of the first principal stress in the local dangerous part of the structure. The local geometric characteristics are precisely measured and compared through finite element simulation to determine the size, shape and sampling angle of the sample. In step S2, a non-contact mechanical testing method is used to obtain the constitutive relationship of the local material after service. Based on the constitutive relationship of the material after service, the full three-dimensional multiphysics simulation results are dynamically corrected. Based on the principle of equal damage, the in-situ test load spectrum is determined. The test load spectrum is any one of the creep damage design load spectrum, creep-low cycle interactive load spectrum, and variable load creep load spectrum. The determination of the in-situ test load spectrum based on the principle of equal damage includes the following steps: Based on the actual service load spectrum, extract the creep duration and low-cycle load cycles; Based on the loads on characteristic parts determined by multiphysics simulation in step S1 under various working conditions, the creep load and low-cycle load cycle of the characteristic parts are determined; When the damage to the specimen under creep load is detected to be greater than the order of magnitude of the damage under low-cycle fatigue load, only the creep damage design load spectrum is considered. Alternatively, when the damage to the specimen under low-cycle fatigue load is detected to be greater than the order of magnitude of the damage under creep load, only the low-cycle fatigue damage design load spectrum is considered. Otherwise, the creep-low-cycle interactive load spectrum is designed based on the damage ratio between the two.

2. The method for assessing the remaining life of high-temperature structures according to claim 1, characterized in that, In step S1, the fixture is designed and verified according to the in-situ test loading scheme to form the geometric structure of the micro in-situ test piece. Then, samples are taken from the batch of parts whose remaining life is to be determined, the deformed layer of the material is removed, and the microstructure of the sample is inspected.

3. The method for assessing the remaining life of high-temperature structures according to claim 1, characterized in that, The loading scheme includes two asymmetric snap clamps (100) and a heating table. The asymmetric snap clamps (100) are arranged on both sides of the test piece (200) along its length. The asymmetric snap clamps (100) are used to provide axial tensile load on the test piece (200). The heating table is used to heat the test area and limit the displacement of the normal phase on the test surface.

4. The method for assessing the remaining life of high-temperature structures according to claim 3, characterized in that, The heating stage includes: Heating ceramic (300) is used to generate heat to heat the test section of the in-situ test piece (200); Two thermocouples (500) are provided. One is connected to the heating ceramic (300) to form a negative feedback channel to ensure accurate application and maintenance of temperature. The other is on the upper surface of the test piece (200) to determine whether the test part has reached thermal equilibrium. Graphite sheet (400) is used to transfer heat to the upper surface of the test piece (200); An elastic base is set in the mounting slot of the asymmetric snap clamp (100) to limit the normal phase displacement of the test piece (200) surface; A heat-insulating baffle (600) is used to protect the electron gun of the scanning electron microscope during high-temperature testing. It is cooled by a serpentine water-cooling channel and its horizontal movement is controlled by a servo motor.

5. The method for assessing the remaining life of high-temperature structures according to claim 1, characterized in that, In step S2, the non-contact mechanical methods include digital image correlation or moiré method. During the in-situ life test of the sample, the strain-time curve, surface displacement distribution and microstructure information of the sample are obtained by non-contact mechanical testing methods.

6. The method for assessing the remaining life of high-temperature structures according to claim 5, characterized in that, In step S2, a high-magnification and large-depth-of-field scanning electron microscope is combined with the experiment to conduct in-situ observation of the microstructure and surface damage evolution of the sample. The scanning electron microscope uses controlled electron beam scanning imaging point by point.

7. The method for assessing the remaining life of high-temperature structures according to claim 1, characterized in that, In step S3, a hybrid model combining a long short-term memory network and a convolutional neural network is used for data processing and feature extraction to identify displacement field features, microstructure features, creep rate and their implicit relationship with remaining lifetime. The long short-term memory network can capture the dependencies in the time series, and the convolutional neural network can automatically extract local spatial features.

Images